Research into programs and policies that can successfully broaden participation in STEM fields shares many of the strengths and weaknesses of the programmatic initiatives directed at the problem. Research has produced much useful information about the design and operation of educational and workplace programs. Yet too often the research findings are poorly integrated into the assessment of existing programs or the development of new ones.

Lessons Learned from Prior Research
Research into the practices and cultures of educational institutions and the workplace can benefit all groups, not just those groups under-represented in STEM fields. As one participant noted, “Underserved groups are like canaries in the coal mine – if an educational institution or workplace has a ‘toxic’ environment, women, persons with disabilities, and under-represented minorities will be the first to suffer.” Research into institutional and interpersonal environments can identify components of a healthy STEM culture and elucidate the characteristics and practices of successful departments, research groups, advisors, and employers.

A wide range of research approaches will be necessary for addressing the many aspects of STEM workforce problems. A necessary component of any program is a robust assessment process to provide feedback on the effectiveness of programs as they are being implemented. Formative assessments that focused on processes can help programs identify what is working and what is not working and can provide feedback that programs can apply in real time. Summative assessments focused on outcomes and products can provide valuable lessons for policymakers, leaders and administrators of other projects. As in the discussion of successful programs, several issues arose repeatedly in discussions of research needs.

The Potential Contributions of the Social Sciences: Many workshop participants pointed toward the value of the social sciences in not only evaluating programs and policies, but also in developing new initiatives. Social scientists are rarely involved in the design or evaluation of programs for increasing participation in the STEM workforce, even though there is a clear need to bring their insights and expertise to bear. Furthermore, the social sciences need to be seen as an integral part of the research enterprise, with similar needs for a diverse and well-trained workforce and with special expertise about the best ways to meet those needs (Levine, Abler, and Rosich, 2004).

For example, social scientists can help identify and address factors that appear to pose barriers to participation in STEM fields, such as institutional and departmental attitudes toward women, persons with disabilities, and under-represented minorities. They can investigate the tensions between meeting community needs in places such as American Indian reservations and training individuals who may leave their communities and thereby reduce the resources available to those communities. They can function as “institutional anthropologists” in examining the social context and organizational arrangement of an educational institution or workplace. They can investigate topics in which little work has previously been conducted — such as the distinct barriers to participation in STEM fields by people with disabilities.
By bridging STEM disciplines, social scientists create an additional pool of qualified individuals to increase the recruitment and retention of more U.S. citizens and especially women and minorities into the STEM workforce. In addition to social scientists, researchers from business schools who study management and organizations can provide valuable insight into institutional change and organizational behavior.

The Importance of Disaggregated Data: Current data often are not collected or presented in enough detail to understand issues and trends for specific groups. Although women, minorities, and people with disabilities share under-representation, the factors causing that under representation may differ from one group to another as well as from one discipline to another. As one participant quipped, “the assumption becomes that all women are white, and all minorities are men.” Data need to be disaggregated by race, gender, disability, field, sector, academic degree, and other characteristics, workshop participants said. Potential similarities among underserved group, such as isolation, tokenism, and undervaluation, need to be investigated, as do potential differences, such as self-confidence, academic experiences, and attrition. Effective solutions then can be tailored to serve the needs of each group.


A significant problem with data disaggregation is that group sizes become too small to analyze effectively or even to discuss, given privacy concerns. Innovative methodological solutions need to be devised to deal with this problem. Longitudinal surveys, case studies, and working with specific individuals all offer possible solutions to small cell sizes. Existing disaggregated data already point toward intriguing results. For example, within the social sciences, sociology has significantly more African American, Hispanic, and women participants than does economics; also in economics a disproportionate number of doctorates are awarded to non-U.S. citizens (Commission on Professionals in Science and Technology, 2000). In general, group experiences by field offer many fruitful research questions — for example, why women are more likely to study the biological sciences than the physical sciences, and whether all Asians – or only Asian men — are disproportionately represented in some fields (National Science Foundation, 2002; Xie and Shauman, 2003). Similarly it is imperative that more research is done to examine how different disabilities are dealt with in the work place. There remains a major gap in our understanding of the impact and accommodations needed for various types of disabilities if persons with disabilities are going to enjoy full participation in the STEM workforce. Researchers need to examine what we know so far and what we need to know about various types of disabilities. Some workshop participants called for more research on persons with non-physical disabilities. We continue to know little about how or if the workplace accommodates workers with disabilities associated with aging. This is an area where new research tools and techniques are needed to allow sufficient disaggregation of data to produce meaningful results specific to particular types of disabilities, while at the same time protecting the confidentiality of individuals involved in the studies.

Strategic Research Areas and Integrated Funding Priorities
Workshop participants identified several features of a more integrated and comprehensive approach to research on the factors that contribute to the choice of a STEM career. The research areas articulated below need to be carried out at both the institutional and individual levels. In addition new methodological approaches and tools need to be developed to facilitate this research.

Research on the beliefs and actions of gatekeepers: The beliefs, attitudes, and practices of faculty and administrative gatekeepers at all levels of education and employment are a critical factor in attracting and retaining STEM students and employees. Similarly, experimental modeling of hypothesized mechanisms for change, collaboration, and negotiation could stimulate the greater participation of under-represented groups in the STEM workforce.

Research on the “culture of science”: A key challenge for attracting people to STEM careers in future years will be improving educational and working conditions in these fields. Long hours, the scarcity of funding for younger investigators in many fields, for women the tension between the biological and tenure clocks, and a lack of rewarding work for graduate and postdoctoral researchers are among many barriers in academe and industry that may discourage participation in STEM careers. An important question to examine is why such factors seem have greater impact on STEM than other professional careers, such as law and medicine (Teitelbaum, 2001). An often-cited problem is the existing practice of science based upon a traditional, male-oriented model that presumes the ability of the researcher to be totally devoted to the scientific endeavor due to a helpmate in the background taking care of children and family concerns. This model no longer works within the current culture of dual career couples who now often share childcare and family responsibilities equally and may deter both men and women from participating in STEM careers. Further understanding of how the structure of institutions and organizations and the practice of science can influence career choice could provide insight into recruiting and retaining a diverse STEM workforce.

Research on individual choices: Investigations of the social and psychological factors contributing to individual choices can reveal factors that appear to pull or push students or employees into or away from STEM fields. Examples include personal and professional self-image, family or community expectations, and existing or apparent tradeoffs between a STEM career and family obligations.

Research on different levels of workforce development: Attention should be focused not just on the participants in the STEM educational and career pathways but on the system within which people operate. In this way, research can emphasize the connections among policies, practices, and people in educational institutions and the workplace. Such research also can examine the important issues of sustainability and scalability.

Research on career patterns and experiences of under-represented minorities and women: Longitudinal, survey, and case study data need to be gathered to determine whether certain career patterns and experiences exist that marginalize these groups and prevent their progress in STEM fields. In particular, data about the under-representation of minorities and women in all STEM fields at top tier research universities need to be gathered to determine the causes behind this under-representation and to suggest remedies.
Research on leadership: Practices in such areas as recruitment, retention, recognition, and promotion can heavily influence advancement to positions of scientific and administrative influence, and this can be an important focus for research on STEM program outcomes.

Research using existing databases: The databases of individual programs should be designed for both research and administrative purposes. Historical data should be preserved for future study rather than being discarded. Ways of funding the creation, preservation and maintenance of these databases must be included in programs.

Evaluation Methodologies and Metrics to Assess Success: While participants were unable to devote much time to the issue of evaluation methodologies and metrics, these issues were referenced in several contexts, most notably exemplary programs. Participants pointed out that only recently has program evaluation received the attention that it deserves. Much of the early federal funding of programs did not include budgetary support for evaluation. Consequently, evidence of program effectiveness was elusive. Although there is more awareness of and inclusion of evaluative components in programs, funding still lags. Many participants called for more funding support for evaluation research (especially involving cultural competency) and for training evaluation researchers, particularly from underrepresented groups.
.